Lyman-alpha Pressure: A Key Feedback Player at Cosmic Dawn ✦
The dynamic role of Lyman-α radiation pressure feedback has remained uncertain due to a lack of multidimensional calculations. This work presents the first two-dimensional radiation hydrodynamics (RHD) simulations using the Lydion code, which implements the M1 moment method for Lyα transfer with dust dynamics and achieves ~100× speed-up compared to Monte Carlo. Modeling star clusters in dense low-metallicity (Z/Z⊙ ≤ 0.01) clouds showed that Lyα pressure strongly boosts outflows and dominates over direct and infrared radiation, producing forces (2–16) Lbol/c and boost factors MF ≈ 10–60. Although leakage in regions of low optical depth, Doppler shifts, and photon destruction weaken the effect, they do not prevent strong pressure buildup in H II regions. Thus, Lyα feedback raises the threshold gas surface density for efficient star formation and is currently missing from almost all galaxy simulations.
Context
For decades, astrophysicists have debated whether radiation pressure from multiple scatterings of the hydrogen Lyman-alpha line emission (Lyα) can influence gas dynamics around massive stars. Early work in this field dates back to pioneers such as Subrahmanyan Chandrasekhar, and in recent years, analytical models and one-dimensional simulations predicted that Lyα feedback could boost radiation pressure by orders of magnitude and dominate in environments poor in cosmic dust. If true, the implications are enormous: from the formation of the first stars and galaxies to the assembly of supermassive black holes at Cosmic Dawn, now actively probed by the James Webb Space Telescope (JWST). However, due to the immense computational cost of Lyα photon transport, no multidimensional simulation has self-consistently included this process, leaving the key question open.
Methods
To fill this gap, the team developed the Lydion code, which implements a novel M1 moment method to solve the Lyα radiation transfer equation in the Fokker–Planck approximation. Unlike traditional Monte Carlo methods that require tracking billions of photons, the M1 approach tracks moments of intensity (mean intensity and flux) on a spatial and frequency grid, achieving a speedup of up to ~100 times. Lydion self-consistently accounts for dust dynamics, gas heating and cooling, chemistry of hydrogen, helium, carbon, and oxygen, as well as feedback from photoionization, direct and infrared radiation pressure. Two-dimensional RHD simulations were performed for compact star clusters (10⁴ M⊙) and isolated massive stars (35 M⊙) embedded in dense (nH ~10⁵–10⁶ cm⁻³) metal- and dust-poor (Z/Z⊙ ≤ 0.01) clouds typical of the early universe. For comparison, cases without Lyα feedback were also modeled.
Results
The results showed that Lyα pressure has a dramatic impact. In simulations without Lyα feedback, clouds quickly collapsed, which would lead to efficient star formation. When Lyα pressure was included, powerful outflows were launched, disrupting the cloud and suppressing collapse. The radial force of Lyα radiation reached impressive values: for star clusters it was ~(2–10)×Lbol/c, and for isolated stars up to 16×Lbol/c, as confirmed by force multipliers MF ~10–60. For comparison, direct radiation pressure contributed only ~0.8 Lbol/c, and infrared pressure less than 0.01 Lbol/c. Even with developed turbulence and the emergence of Lyα photon leakage channels through low optical depth regions, the pressure force decreased only modestly (by ~2 times), remaining dominant. Moreover, Lyα pressure exceeded gas pressure over a large part of the H II region, especially near the ionization front. Simulations of isolated stars in dust-poor conditions (D/D⊙ ~10⁻³) demonstrated that accounting for Lyα photon destruction (2p→2s transitions) reduces the force multiplier from ~100–300 to ~35–60, but does not negate the overwhelming dominance of Lyα feedback over direct feedback.
Implications
These results fundamentally change our understanding of star formation regulation in the early universe. They show that Lyα feedback is not just a correction but the dominant pre-supernova feedback mechanism in dense, dust-poor environments. Its inclusion is essential for realistic simulations of the formation of the first stars, nebulae, and galaxies. In particular, Lyα pressure can explain how massive star clusters, like early globular clusters, lose gas and halt further growth. Ignoring this process in modern cosmological simulations casts doubt on their predictions for star formation efficiency, initial mass function, and escape of ionizing photons—key parameters for interpreting data from JWST.
Future development
Future development of this topic will require both expanded parameter studies with Lydion (varying cloud mass, metallicity, star formation rate) and the creation of accelerated Lyα transfer methods for three-dimensional simulations. A priority task will be the development of subgrid models for Lyα feedback in large-scale cosmological simulations, similar to how other feedback mechanisms are already parameterized. Recent analytical solutions verified by Monte Carlo methods could serve as a basis for such models. It is also critically important to incorporate more complex dust physics—its growth, destruction, and dynamics—which significantly affects Lyα photon absorption and thus the feedback strength.
Impact
The results will impact several areas of astrophysics: the theory of star formation in extreme conditions, modeling of the first galaxies and their observable properties (spectra, Lyα luminosities), and cosmological simulations that use JWST to test the ΛCDM model.
Next steps
Near-term steps include extending Lydion simulations to three dimensions, integrating a realistic star formation model, and comparing with new JWST observational targets, such as compact star clusters and direct-collapse black hole candidates.
Key open problems
This research is directly connected to several unsolved problems in physics: how the first stars (Population III) formed and why their typical masses might be lower than predicted; what limits the growth of supermassive black holes at early epochs; and what is the source of the escaping ionizing radiation that ended the epoch of reionization. Lyα feedback could be the missing link responsible for self-regulating these processes, and its inclusion may resolve long-standing tensions between theory and observations.
🎯 The Lyman-alpha line is the brightest spectral line of hydrogen, and in optically thick environments, a single photon can scatter millions of times before escaping the cloud. With each scattering, it transfers momentum to the gas, and in a dust-poor medium, this creates a pressure comparable to the radiation pressure from all other stellar photons combined!
🎬 The idea of using radiation pressure for propulsion is not new: in Arthur C. Clarke's novel 'Rendezvous with Rama,' a solar sail harnesses light pressure. Lyα feedback is like a natural 'super-sail,' where multiple scattering multiplies the force by tens of times, but here it can both accelerate and disrupt gas clouds around stars.
Key numbers
- Lyα pressure force: 2–16 Lbol/c
- force multipliers MF: 10–60
- initial optical depth of cloud: up to 10^23 cm⁻²
- speedup of RHD code compared to Monte Carlo: ~100 times
- mass of star cluster in simulation: 10^4 M_sun